Multi-scale Chlorophyll Fluorescence And Photosynthetic Response Mechanisms To Drought Stress And Integration Methods With Crop Model

In the context of global warming and decreasing precipitation,drought disasters pose a serious threat to regional crop yields.Crop models can be used to predict and evaluate the impact of climate change on crop yields.However,because crop models require a large amount of field measurement data for parameter calibration,they are mostly used for field-scale simulation,and there is considerable uncertainty in simulating at the regional level.Solar-induced chlorophyll fluorescence（SIF）is a new remote sensing indicator that can directly monitor the photosynthetic productivity of regional crops.Combining SIF with crop models can help constrain the photosynthetic simulation process of crop models,thereby improving the regional yield simulation ability of crop models.However,under stress conditions,the linear relationship between SIF with scale effects and vegetation photosynthesis may become decoupled,and there may be significant issues with directly using SIF for photosynthetic estimation or coupling with crop models under stress conditions.Therefore,based on the chlorophyll fluorescence and photosynthetic response mechanisms under drought stress,this study coupled fluorescence algorithms with the Agricultural Production Systems s IMulator（APSIM）crop model as the core,and conducted simulations and analyses from four aspects:the photosystem scale,leaf scale,field scale,and regional scale.At the photosystem scale,the fluorescence emission spectra were reconstructed,and the fluorescence quantum efficiency of photosystem I（PSI）and photosystem II（PSII）were separated,providing a theoretical basis for modeling photosystem fluorescence.At the leaf scale,changes in the chlorophyll fluorescence spectrum,gas exchange parameters,and Pulse amplitude modulated（PAM）fluorescence parameters of maize leaves under different degrees of drought stress were analyzed based on indoor drought experiments,which can be used to improve the photosynthesis-fluorescence modeling under stress conditions.At the field scale,the simulation of canopy SIF escape probability in farmland and the analysis of its influencing factors were carried out.Based on existing theoretical knowledge and research results under stress conditions,the APSIM-SIF model was developed and improved to enhance its applicability under drought stress.At the regional scale,the APSIM-SIF model was applied to simulate SIF in the Midwest region of the United States,validated using TROPOMI SIF data,and used to simulate crop yields in the region by constraining the APSIM-SIF model with remote sensing SIF data.The main research results and conclusions of this paper are as follows:（1）Mechanism for converting full-spectrum SIF emitted by leaves to single-band SIF at the top of the canopyBy using the fluorescence excitation matrix to calculate the SIF band conversion factor（ε）and the spectral invariance theory to estimate the fluorescence escape probability(f_esc),the full-spectrum SIF emitted by leaves can be converted to single-band SIF at the top of the canopy.εand f_esc can respectively represent the physiological state of the leaves and the structural characteristics of the plant canopy.εcan be calculated using biochemical parameters inputted into the FLUSPECT model through the excitation-fluorescence matrices calculation.The f_esc of near-infrared SIF can be expressed as a function of directional reflectance in the NIR region(R_NIR),normalized difference vegetation index（NDVI）,and the fraction of absorbed photosynthetically active radiation(f_APAR).Sensitivity analysis results show that chlorophyll concentration is the main driving factor ofεin the NIR region,which can explain more than 80%of the variation inε.Regional simulation results show that f_esc remains relatively constant during the crop growth peak period.（2）Reconstruction of the photosystem fluorescence emission spectrum and separation of the fluorescence quantum efficiency based on the FLUSPECT modelUsing the singular value decomposition（SVD）method to separate sample weights and singular vectors of the fluorescence emission spectrum,the decomposition results show that the first three principal components can explain more than 90%of the spectral variability.Based on non-photochemical quenching（NPQ）,net photosynthesis rate,and the first three principal components,the fluorescence spectrum can be reconstructed.Integrating the reconstructed fluorescence emission spectrum into the FLUSPECT model improves its accuracy in simulating fluorescence spectra under light,temperature,and CO₂ environments（R²>0.9）.The conversion coefficient between PSII fluorescence quantum efficiency and the PSII fluorescence yield measured by PAM is determined to be 0.87 using linear regression.Under short-term changes in light intensity,the PSII fluorescence quantum efficiency reaches its highest value of 0.012 in weak light and then decreases with increasing light intensity,while the PSI fluorescence quantum efficiency remains stable and contributes up to 1/3 of the total fluorescence quantum efficiency under strong light.（3）Improving the linear relationship between chlorophyll fluorescence and photosynthesis under drought stress based on leaf-level experimentsUnder drought stress,the linear relationship between chlorophyll fluorescence（ChlF）and photosynthesis is decoupled at the leaf level.As the soil moisture content decreases,the changes in photosynthesis rate under different light intensities show a consistent trend with stomatal conductance and electron transfer rate,all exhibiting a distinct threshold of soil water content.However,the decrease of ChlF with decreasing soil moisture content is not significant,and even in the red band,ChlF may increase.Therefore,under sustained drought stress,the linear relationship between ChlF and photosynthesis rate is very low（R²<0.2）.Based on the electron transfer rate formula,introducing the degree of openness of the PSII reaction center（q_L）parameter can significantly improve the estimation ability of constrained chlorophyll fluorescence（ChlF×q_L）for photosynthesis under stress conditions（R²>0.9）.（4）Simulation of canopy fluorescence escape probability and influencing factors in farmland based on the canopy radiative transfer modelsUsing the one-dimensional SCOPE model to study the effect of observation direction on canopy fluorescence escape probability,the results show that directional fluorescence escape probability is greatly influenced by leaf inclination angle distribution function（LIDF）,observation angle,and leaf area index（LAI）,while hemispherical fluorescence escape probability is mainly related to LAI and LIDF.Assuming that the LIDF of the farmland canopy is spherically distributed,the accurate simulation of hemispherical fluorescence escape probability can be achieved by LAI and the scattering and absorption coefficients of different bands（R²>0.95）.The directional fluorescence escape probability can be estimated based on R_NIR·NDVI/f_APAR（R²=0.89）.By setting different crop planting scenarios with different plant and row spacing in the DART model for three-dimensional radiative transfer simulation,the results show that under high planting density,soil reflectance has a low impact on fluorescence escape probability,while under low planting density,soil reflectance has a significant impact.Compared with plant spacing,row spacing has a greater impact on canopy fluorescence escape probability.（5）Construction and improvement of the APSIM-SIF modelWe developed an SIF module within the APSIM model framework to couple crop canopy gross primary productivity（GPP）with top-of-canopy SIF(SIF_toc).The APSIM-SIF model was validated at two agricultural sites in the Midwest United States and was able to predict over90%of the variation in GPP,aboveground biomass,and leaf area index（LAI）.It also captured well the seasonal variation in SIF_toc（R²=0.84）and GPP（R²=0.81）at a maize irrigation site in China.For grid points where crops occupied more than 70%of the area,simulated SIF_tocexplained over 75%of the variability in satellite-observed SIF.We improved the photosynthesis-fluorescence coupling system in the APSIM-SIF model based on the Mechanistic Light Response（MLR）method,and downscaled the leaf-level simulation to the photosystem-level simulation.The results of validation showed that the improved APSIM-SIF model significantly improved the simulation accuracy of total canopy SIF（R²>0.9,RRMSE<20%）and enhanced the response to drought stress.Based on the MLR method,the APSIM-SIF model simulated crop yield with a high correlation with the county-level crop yield of the National Agricultural Statistics Service,with the highest correlation being 0.78 and the lowest simulation error being 748 kg ha^-1.The conclusions of this study are as follows:Using reconstructed fluorescence emission spectrum at the photosystem scale helps improve the simulation of leaf fluorescence spectra in different environments.Introducing the q_L parameter in fluorescence-photosynthesis modeling at the leaf scale significantly enhances the estimation ability of photosynthesis in stressful environments.At the canopy scale,LAI and LIDF have a significant impact on early canopy fluorescence escape probability,while fluorescence escape probability remains relatively stable during the crop’s peak season.At the regional scale,the APSIM-SIF model accurately simulates the spatiotemporal variations of regional SIF.Downscaling leaf fluorescence simulation to the photosystem scale and introducing the q_L parameter contribute to enhancing the sensitivity of the APSIM-SIF model to drought responses and improving its capability to simulate regional crop yields.